作者: Yong Liang , Long He , Xinyu Cao , Zuo-Jun Shen
关键词:
摘要: In this paper, we study the optimal control problem for demand-side of smart grid under time-varying prices with general structures. We assume that users are equipped appliances allow delay in satisfying demands, and one central controller makes energy usage decisions on when how to satisfy scheduled demands. formulate a dynamic programming model problem. The deals stochastic demand arrivals schedules demands based their own allowable delays, which specified by users. However, encounters “curses dimensionality” some other difficulties, thus is hard solve. develop decentralization-based heuristic first, also propose an approximation approach Q-learning. Finally, conduct numerical studies testing simulation results show both Q-learning decentralization approaches work well, but they have advantages disadvantages different scenarios. Lastly, conclude paper discussions future extension directions.